Feedback on EmberZNet 8.0.0

[Nerivec]

Resolved

• [Fixed in 8.0.2] The new "rx packet info" struct used in handlers in place of LQI, RSSI, etc params appears to return improper values for the last 3 variables.
  • Last hop LQI is seen as a zero value after reset that increases with every command sent to the device, so clearly, it's a value somewhere, just not the right one (though sometimes it's just 255).
  • Last hop RSSI is seen as random values (positive, negative, zero).
  • Last hop timestamp is new, so, not sure if it's correct (seems a bit strange though).
  • Excerpts from two INCOMING_MESSAGE_HANDLER received in succession:
    • dec90000000000000000ff08060000000000
    • dec90000000000000000ff08070000000000
  • The values appear to be correct in a (very) few random ZDO responses.
• [Appears fixed in 8.0.1] The NCP appears to be a lot better at handling stress with 7.4.x than 8.0.0. (same hardware, same host app, only firmware version changed)
  • I'm seeing ERROR_EXCEEDED_MAXIMUM_ACK_TIMEOUT_COUNT 100% of the time after multiple SEND_UNICAST in rapid succession. It handles the first several without problem, but eventually it seems to miss/drop a reply to SEND_UNICAST (never seen in the host app) despite apparently registering it as sent (since it times out).
  • I can increase the amount of unicasts ten fold and still not see this issue with 7.4.x.

• SDK code for launch standalone bootloader command indicates it was switched to sl_status_t, but payload received from firmware indicates it did not (still 1 byte).

Note: Will keep this updated if new issues are discovered or more information can be provided.

[Hedda]

Also see related #11

[nkljajic]

Add call to sli_zigbee_fill_rx_packet_info() early in sli_zigbee_call_incoming_message_handler()

[Hedda]

FYI, latest Silabs official documentation seems to now have been updated to reflect changes in EZSPv14 (and EmberZNet 8.0.x.x):

https://docs.silabs.com/zigbee/8.0.1/zigbee-stack-api/ember

https://docs.silabs.com/zigbee/latest/zigbee-stack-api/ember